Trump-Stock-Tips/trump-bot.py

# -*- coding: utf-8 -*-

import re
import string
import requests
import ast
import os
import random
import twitter
import json
import time
from datetime import datetime
import gspread
from oauth2client.service_account import ServiceAccountCredentials
from stopwords import stop_words
from companies import all_companies


def find_companies(word, dictionary):
    ''' 
    search dictionary of company names for a match with word
    returns array of company names
    '''
    search_ex = "(^|\s)" + word + "(?=,|\s)"
    results = []
    for company in dictionary:
        if re.search(search_ex, company['Name'], re.IGNORECASE) is not None:
            results.append(company)
    return results

def find_all_companies(text):
    '''
    Returns an array of matched companies for each word in the text
    [[word, [company, ..., company]],[word, [company, ..., company]]]
    '''
    for c in string.punctuation:
            text = text.replace(c,"")
    text = " ".join(text.split())
    words = text.split(" ")
    results = []
    for word in words:
        # don't process digits
        if word.isdigit() == False:      
            if word not in stop_words and word.lower() not in stop_words and word.title() not in stop_words:
                companies = find_companies(word, all_companies)
                if companies:
                    results.append([word, companies])
    return results

def test_sentiment(text):
    '''
    returns sentiment analysis label: pos, neg or neutral
    using http://text-processing.com/docs/sentiment.html
    '''
    r = requests.post("http://text-processing.com/api/sentiment/", data = {"text": text})
    sentiment = ast.literal_eval(r.text)

    pos_prob = sentiment['probability']['pos']
    neg_prob = sentiment['probability']['neg']
    neutral_prob = sentiment['probability']['neutral']
    if neutral_prob > 0.9:
        label = "neutral"
    elif pos_prob > neg_prob:
        label = "pos"
    elif neg_prob > pos_prob:
        label = "neg"
    return label

def exclamation(n):
    ''' Returns a string of n random ?s and !s'''
    exclamation = ""
    for i in range(n):
        exclamation += random.choice("?!")
    return exclamation

def log_tip(spreadsheet, tip, symbol, name):
    list_of_lists = spreadsheet.get_all_values()
    next_row_nb = len(list_of_lists)+1
    time_stamp = str(datetime.now())

    row_values = [time_stamp, tip, symbol, name]

    for value in row_values:
        row_cell = row_values.index(value) + 1
        spreadsheet.update_cell(next_row_nb, row_cell, value)

def tweet_to_tip(tweet, spreadsheet):
    '''
    Returns a stock market tip from an input tweet (if any matches are found)
    logs the tip to a google spreadsheet using 
    '''
    names = ["He", "Trump", "POTUS"]
    name = random.choice(names)
    hits = find_all_companies(tweet)
    if hits:
        chosen_hit = random.choice(hits)
        sentiment = test_sentiment(tweet)

        word = chosen_hit[0].strip()
        stocks = chosen_hit[1]
        random.shuffle(stocks)
        Symbol = "$" + stocks[0]['Symbol']
        Name = stocks[0]['Name']
        if sentiment == 'pos':
            excl = exclamation(3)
            intro = name + " loves " + "\"" + word + "\""
            tip = "BUY "
            stock_tip = tip + Symbol + " ↗️ " + Name
        elif sentiment == 'neg':
            excl = exclamation(3)
            intro = name + " hates " + "\"" + word + "\""
            tip = "SELL " 
            stock_tip = tip + Symbol + " ↘️ " + Name
        elif sentiment == 'neutral':
            excl = exclamation(1)
            intro = name + " says " + "\"" + word + "\""
            tip = "WATCH " 
            stock_tip = tip + Symbol + " 🔎 " + Name
        log_tip(spreadsheet, tip, Symbol, Name)
        tip_tweet = intro + excl + "\n" + "TIP: " + stock_tip
        return tip_tweet
    else:
        print "no hits"
        pass

def twitter_api_auth(is_prod):
    ''' authenticate twitter api '''

    if is_prod:
        api = twitter.Api(consumer_key=os.environ.get('consumer_key'), consumer_secret=os.environ.get('consumer_secret'), access_token_key=os.environ.get('access_token_key'), access_token_secret=os.environ.get('access_token_secret'))
    else:
        from twitter_creds import twitter_creds
        api = twitter.Api(consumer_key=twitter_creds['consumer_key'], consumer_secret=twitter_creds['consumer_secret'], access_token_key=twitter_creds['access_token_key'], access_token_secret=twitter_creds['access_token_secret'])

    return api

def gsheets_api_keys(is_prod):
    ''' Returns credentails for Google API Oauth '''

    scope = ['https://spreadsheets.google.com/feeds']
    
    if is_prod:
        with open('gsheets_creds.json', 'w') as file:
            json.dump(ast.literal_eval(os.environ.get('gsheets_creds')), file)
        gsheets_credentials = ServiceAccountCredentials.from_json_keyfile_name('gsheets_creds.json', scope)
        os.remove('gsheets_creds.json') 

    else:
        gsheets_credentials = ServiceAccountCredentials.from_json_keyfile_name('trump-bot-0b1e240571d7.json', scope)

    return gsheets_credentials

def main():
    import sys;
    reload(sys);
    sys.setdefaultencoding("utf8")

    is_prod = os.environ.get('IS_HEROKU', None)
    twitter_api = twitter_api_auth(is_prod)
    gsheets_credentials = gsheets_api_keys(is_prod)

    timer = 60*10

    while True:
        print "--- start ---"
        
        # auth runs out after a while and throws error
        gsheets_api = gspread.authorize(gsheets_credentials)
        id_worksheet = gsheets_api.open("trump-bot").sheet1
        logs_worksheet = gsheets_api.open("trump-bot").worksheet('logs')
        
        last_id = int(id_worksheet.acell('A1').value)
        print "last_id:", last_id
        
        tweets = twitter_api.GetUserTimeline(screen_name='realdonaldtrump', since_id=last_id)
        
        tweet_queue = []
        if tweets:
            for tweet in tweets:
                parsed_tweet = json.loads(str(tweet))
                tweet_queue.append([parsed_tweet['id'], parsed_tweet['text']])

        if tweet_queue:
            # reversed list to process most recent tweet last and prevent last_id mess
            for tweet in reversed(tweet_queue):
                tweet_id = tweet[0]
                tweet_text = tweet[1]
                print "new tweet:", tweet_text, tweet_id
                try:
                    top_tip = tweet_to_tip(tweet_text, logs_worksheet) #+ "\n" + tweet_link
                    print "tweeting: " + top_tip
                    status = twitter_api.PostUpdate(top_tip)
                except:
                    pass

                id_worksheet.update_acell('A1', str(tweet_id))
                print "last_id updated to: ", tweet_id, "\n"

        print "--- done --- \n"
        time.sleep(timer)

if __name__ == '__main__':
    main()